How to use your phone camera to identify skin cancer

Your smartphone's camera is useful in so many different ways. You depend on it for Snapchat, Instagram ... and identifying skin cancer?

Melanoma, the deadliest form of skin cancer, is expected to cause more than 10,000 deaths in the U.S. alone in 2016. Researchers are now hard-pressed to find a new way to catch the disease and others like it in the earliest stages. That's where that handy, ubiquitous iPhone camera can help, according to new research.

In an IBM Research Blog post, Dr. Noel Codella outlines a means of identifying markers of melanoma via skin image analysis that might be available to doctors and patients in the future.

The methodology for home diagnosis via smartphone is relatively simple (at least in theory): When someone finds a questionable spot on their skin, they use their handset's camera to take a picture of the lesion and submit the image to be assessed by an analytics service, which can recognize and reliably identify the characteristics of disease.

In practice, though, it's much more complicated than that.

We've seen this type of system put to the test before in standalone apps, but those programs were woefully inadequate at best, resulting in a dreadful 93 percent failure rate in some instances.

But that was all the way back in 2013. Now, the IBM team is employing much more powerful tools to improve the accuracy of computer image analysis.

The key to the success of this project hinges on two factors. The first is the widespread use of Dermascopes, which are devices that can be attached to smartphone cameras to optimize photos of lesions for analyzation. The second (and more important) factor is the development of a massive database containing images of cancerous skin spots. The database is accessed using IBM's machine learning, computer vision and cloud computing capabilities to develop the means to consistently identify cases of melanoma through technology.

This system was built through a partnership with Memorial Sloan Kettering Cancer Center and the International Skin Imaging Collaboration (ISIC). Preliminary research published in 2015 using ISIC data outlined the concept of using computer vision tech to identify disease and boasted promising early returns — but it was still very user intensive and limited. Most notably, there was "no direct comparison to the diagnostic performance of human experts," so there was no way to judge the results generated by the image analysis against the findings of an experienced dermatologist.

After two further years of development and expansion of the image database, the method has become three times more effective. Importantly, the it was helped along by comparing its results against those of eight specialists — and now the tech can do as well at recognizing disease as the experts that contributed to the research.

The IBM team will publish a report on the state of the project in a 2017 issue of the IBM Journal of Research and Development, which can be already be accessed online.

Does this mean that diagnosing melanoma is now as simple as snapping a picture and submitting it to the database? According to Codella, we're still not there yet.

IBM isn't developing a sure-fire method to diagnose disease — instead, this should be thought of as another asset for doctors in their practice, Codella said.

"If you have a lesion, a doctor has a variety of tools to diagnose that lesion," he told Mashable. "We’re looking to augment [those] with another tool to help catch that disease and to even assist general practitioners to be more helpful for their patients."

And like any other scientific development, the work will require rigorous peer review before being accepted by the community. Thankfully, the openness of the database will help to speed that process along.

"We’re using datasets for our studies that are open and in the public domain," Codella said. "We evaluated on a test set that is publicly available. People can vet our work and confirm and evaluate based on that public data."

Still, one might assume that since the computer can now match the abilities of the experts (and will only improve with more data) it will be ready for widespread use soon. But again, Codella stresses that it still has a long way to go.

"This is early work," he said. "We’re actively working with our partners to determine that point. We’re not there yet."